Silverman R H, Coleman D J, Rondeau M J, Woods S M, Lizzi F L
Margaret M. Dyson Vision Research Institute, Cornell University Medical College, New York, NY 10021.
Retina. 1993;13(1):69-74. doi: 10.1097/00006982-199313010-00015.
A system has been developed to determine intraocular tumor volume, a characteristic that has been estimated in previous studies from linear dimensions using a variety of models. Volume was determined by tracing tumor boundaries in sequential, parallel ultrasound scans, adding the areas and multiplying by the interslice interval. The in vitro accuracy of the technique was within +/- 2%. The mean difference between volumes determined from replicate in vivo scans of intraocular tumors was 4.3%. Serial scan volumes were, on average, 19% smaller than volumes computed from an ellipsoidal model and 13% smaller than area-rotational volumes. Differences of as much as 50% were observed between serial scan volumes and volumes computed with these models. The results indicate that methods based on either linear measurements or the tumor area in a single cross-section will tend to both systematically overestimate tumor volume and suffer from unpredictable, nonsystematic errors.
已开发出一种用于确定眼内肿瘤体积的系统，此前的研究曾使用各种模型根据线性尺寸来估算这一特征。通过在连续、平行的超声扫描中描绘肿瘤边界，将各层面的面积相加并乘以层间距来确定体积。该技术的体外准确性在±2%以内。对眼内肿瘤进行重复活体扫描所确定的体积之间的平均差异为4.3%。连续扫描得到的体积平均比根据椭球体模型计算出的体积小19%，比面积旋转体积小13%。连续扫描体积与用这些模型计算出的体积之间观察到的差异高达50%。结果表明，基于线性测量或单个横截面中肿瘤面积的方法往往会系统性地高估肿瘤体积，并存在不可预测的非系统性误差。
